Trust Debt
The problem isn’t that something went wrong. It’s how it was built to go wrong.
There’s a particular feeling I want to describe.
It’s the one you get when you read about a corporate failure — a real one, with real consequences for real people — and your first reaction isn’t outrage. It’s recognition.
Last month, a joint investigation by Svenska Dagbladet and Göteborgs-Posten revealed that footage captured through Meta’s Ray-Ban smart glasses — from bedrooms, bathrooms, and living rooms — was making its way to data annotators at a contractor facility in Nairobi, Kenya. Workers described reviewing intimate moments that users had no idea were being captured, let alone watched by strangers. When workers raised concerns, they were told: “Don’t ask questions, or you’re gone.”
I read this and recognised it immediately. Not because I’ve seen this specific failure. Because I’ve seen the conditions that produce it — in rooms I sit in every week.
In engineering, there’s a concept called technical debt. The shortcuts you take to ship faster. Decisions that feel pragmatic in the moment but accumulate quietly, invisibly, until the day the cost arrives all at once.
Trust works the same way. I call this trust debt. And unlike technical debt, it never shows up in your backlog.
Trust debt doesn’t begin with a villain making a calculated choice to harm users. It begins much more ordinarily. It begins in a meeting.
A commercial stakeholder presents numbers. Conversion is up. Adoption is growing. The slide is confident, the metrics are real, and somewhere in the margin, if it appears at all, is a trust metric. A fraud signal. A customer impact figure. And when it does appear, it tends to arrive as a percentage. A ratio. A number cleanly separated from the person it represents.
The question that doesn’t get asked in that meeting is the one that matters most: who is behind that number? What actually happened to them?
I’ve watched this dynamic play out at leadership levels where people genuinely care — where, when it’s raised, the concern is acknowledged sincerely. And then the meeting ends. The roadmap continues. The commercial momentum carries forward. Not out of malice. Out of structure. The unidirectional pull of growth targets is simply stronger than the diffuse human weight of trust exposure — until the day it isn’t.
That’s the moment trust debt comes due.
Here’s what the accumulation looks like, step by step.
A PM asks about consent flows. Legal signs off. The PM moves on. Nobody asks what a real person — reading nothing, clicking through quickly — would actually expect to happen to their data.
An engineer flags that the anonymisation tool has a ninety-something percent accuracy rate. The team ships. Nobody asks what happens in the remaining cases — under difficult lighting, at scale, across millions of people.
A vendor relationship gets flagged as a procurement question. Procurement handles it. Nobody asks what happens inside that vendor’s building once the data arrives — what the workers see, what they’re allowed to say, what happens when something feels wrong.
Each decision is defensible in isolation. None of them feels like an ethical failure in the moment. That’s what makes trust debt dangerous. It doesn’t announce itself. It accumulates through the ordinary functioning of a product organisation that has decided — not through bad intent but through bad structure — that trust is someone else’s job.
This pattern is not unique to Meta, and it is not unique to tech.
A food platform lists restaurants that don’t exist, processes orders that can’t be fulfilled, and issues vouchers to cover the gap — optimising for order volume while fake supplier and fake customer collusion hollows out the business from inside. Across food delivery and ride-hailing, riders sell verified accounts to strangers who take up the job — the dashboard shows active supply, while the person arriving at a customer’s door is unknown, unvetted, and in some cases, the source of financial, physical, or personal harm. A payments platform onboards millions of users under what regulators later called, in the fine, a “growth at all costs” mentality — KYC checks stripped back to remove friction, conversion climbing, trust exposure invisible until a $504 million penalty made it visible.
Three industries. Three dashboards that showed the right numbers. Three trust debts that were always there — just not yet due.
There is one thing I have come to believe, from sitting on the side of the table that carries these questions every day.
Metrics without soul don’t protect anyone.
A fraud rate of 0.3% is a manageable number until you sit with what it represents — a person who trusted your platform with something real, and had that trust broken. A conversion rate is a clean number until you ask who converted, under what impression, into what they didn’t fully understand.
The organisations that treat trust as infrastructure don’t do so because they are more ethical in the abstract. They do so because someone, somewhere in the building, insists on telling the full story. Not just the metric. The person behind the metric. Not just the conversion number. The customer behind the conversion number.
That insistence — quiet, persistent, present in every meeting where it’s easier to let it slide — is what separates the companies that build trust from the companies that perform it.
Every product team should carry a nightmare customer story. Not a percentage. A real account — what happened, to whom, and what the numbers looked like on the dashboard while it was happening. That story is what you bring to the meeting where the confident slide is presented. It is the other side of the coin. If you don’t have it, you are not measuring trust. You are measuring the absence of a headline.
The bill always arrives. The only question is who pays it — and whether you knew it was coming.
I work in marketplace integrity and fraud prevention. The questions in this piece are not abstract for me — they sit at the centre of what my team navigates every day. When trust fails at scale, the consequences are measurable, costly, and in most cases, entirely avoidable. That’s what makes them worth talking about.
This is the first piece in The Integrity Brief — a publication on trust, integrity, and what it costs when organisations get it wrong.